About Me

I'm the technical lead for Amazon Currency Converter for Sellers (ACCS), a global payments and monetization platform. My current focus is a GenAI agent-based initiative for enterprise accounting onboarding, a process that can take up to six months. Instead of one large agent, it uses narrow, bounded agents and deliberately placed human checkpoints, aiming to bring that time down toward three weeks. The idea behind it applies well beyond accounting: map where the judgment in a workflow actually lives, automate around it, and keep people where a wrong call is expensive to undo.

The thread goes back further. At NASA Ames Research Center I built software for the Tailored Arrival Manager, an autonomous, trajectory-based air traffic system developed with Boeing and the FAA. My Master's research in Computer Science at San Jose State University, on large-scale optimization for emergency wireless networks, led to papers at IEEE and ACM CIKM, and I recently co-authored Agentic AI for Autonomous Decision Systems (CISCom 2026).

I believe the only way to build fast is to build right, so I care as much about the architecture, evaluation, and review practices around an AI system as about the model inside it. Beyond my day job, I write about engineering judgment and GenAI agents, speak and serve on committees at technical conferences, judge hackathons, and mentor engineers. Away from the keyboard, you'll usually find me outdoors.